Seminar

1. Anna Ilina

   Methods for intelligent text data analysis and web tools for information processing and visualization

   Abstract:

   The results are presented on the development and application of advanced methods for processing and intelligent analysis of textual data in scientific and technical domains, as well as on the design of efficient web-based tools for data processing and visualization. In particular, the results of research in the field of semantic analysis and named entity recognition are described. Additionally, solutions are proposed for monitoring resource utilization within the DIRAC distributed infrastructure and ensuring the security of network connections to the internal services of JINR.

2. Daria Zrelova

   Intellectual Analysis of Scientific and Technical Information Based on Patent Documentation

   Abstract:

   This presentation focuses on the intelligent analysis of scientific and technical information, using patent documentation as an example. Patent data represents a relevant source of information on long-term technological trends and the practical implementation of innovations. However, analyzing patents presents certain challenges due to their specific structure, which prioritizes legal protection of the invention rather than a detailed explanation of its essence. This work explores the specifics of patent information analysis, emphasizing modern approaches and methods for extracting valuable insights from patent data to identify promising avenues for technological development. In particular, the study analyzes and examines in detail data related to classifications, dates, languages of submission, authors, owners, and other relevant fields. To conduct semantic analysis, a document corpus was created from patent abstracts, and a corresponding vocabulary was defined. These resources were used to train a neural network language model based on Word2Vec. Semantic analysis is crucial for analyzing the textual content of patents, extracting key terms, phrases, and concepts that characterize each patent, and supporting their classification.

Сonnecting to MTS Link.
Information about the seminar and the link to join are available at Indico.